US20020026298A1 - Method for analyzing a fabrication machine cluster using-genetic algorithms - Google Patents

Method for analyzing a fabrication machine cluster using-genetic algorithms Download PDF

Info

Publication number
US20020026298A1
US20020026298A1 US09/200,330 US20033098A US2002026298A1 US 20020026298 A1 US20020026298 A1 US 20020026298A1 US 20033098 A US20033098 A US 20033098A US 2002026298 A1 US2002026298 A1 US 2002026298A1
Authority
US
United States
Prior art keywords
dna
code
yield
codes
fabrication
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US09/200,330
Inventor
Meng-Lin Yeh
Jasmine Wu
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
United Microelectronics Corp
Original Assignee
United Microelectronics Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by United Microelectronics Corp filed Critical United Microelectronics Corp
Assigned to UNITED SEMICONDUCTOR CORP. reassignment UNITED SEMICONDUCTOR CORP. ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: WU, JASMINE, YEH, MENG-LIN
Assigned to UNITED MICROELECTRONICS CORP. reassignment UNITED MICROELECTRONICS CORP. ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: UNITED SEMICONDUCTOR CORP.
Publication of US20020026298A1 publication Critical patent/US20020026298A1/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion

Definitions

  • This invention relates to semiconductor fabrication, and more particularly to a method using a logic algorithm for analyzing a consequent effect of a fabrication machine cluster.
  • one batch which includes several wafers, needs to pass several fabrication machines in different fabrication processes, such as deposition, etching, or sputtering machines.
  • Each fabrication machine has its own performance and its own series number.
  • a same fabrication process may use same-type fabrication machines provided by different manufacturers. Since semiconductor fabrication needs very high precision and highly reduced dimension, discrepancy of fabrication performance between each of the same-type different machines is strictly required to be consistent in a certain small range.
  • its fabrication performance may also drift away after a period of operation time. All above fabrication variances may cause a fabrication failure due to a consequent effect, which is an accumulated effect of the fabrication variances in each machine used for each fabrication process. The yield rate is therefore reduced. A poor yield rate may particularly occurs at a certain fabrication machine cluster, which includes several machines sequentially used in each step of fabrication processes. Each machine has its own series number to be recorded for the fabrication machine cluster.
  • the analysis method can handle a large amount of data for a complicate fabrication cluster so that an actual reason causing a poor yield rate in a certain fabrication machine cluster can be precisely determined.
  • a modification of on the fabrication machine cluster having low yield rate can be precisely done to increase its yield rate.
  • an analysis method for analyzing a consequent effect of a fabrication cluster includes providing a data base of high-yield DNA codes.
  • a DNA code represents a sequential machine number of a fabrication machine cluster.
  • the data base of the high-yield DNA codes include several DNA codes, all of which have high yield rate.
  • a low-yield DNA code is also provided to be analyzed.
  • An “AND” logic operation preferably is performed to compare the low-yield DNA code with one of the high-yield DNA codes from the data base.
  • Each DNA element code in the DNA codes represents one used fabrication machine in a specified fabrication process.
  • a resulting element code is indicate by “1” for this specified fabrication machine. If they are different, this resulting element code is indicated by “0”. After the low-yield DNA code is completely compared, for example, at least one of the resulting element code is indicated by “0”. A modification process is performed to replace the resulting element having “0” with one of other same-type machines in its specified fabrication process.
  • the analysis method of the invention for analyzing the consequent effect is using a logic algorithm to determine fabrication machines, which may cause the poor yield rate.
  • the analysis method can analyze a large amount of data with a complicate fabrication machine cluster without missing any possible factor causing the poor yield rate.
  • FIG. 1 is a flow diagram of fabrication processes for one batch passing several fabrication machines.
  • FIG. 2 is an AND logic operation algorithm of an analysis method for a consequent effect, according to a preferred embodiment of the invention.
  • each batch has it own log including a history of the batch performed in a certain fabrication machine at a certain step.
  • Each machine is indicated by, for example, an element number or an element code. All the sequentially used fabrication machines form a fabrication machine cluster.
  • the fabrication machine cluster contains those element codes, which form a DNA code like a biologic DNA structure.
  • One DNA code represents one fabrication machine cluster or one history of the batch passing fabrication machines in the whole fabrication process.
  • Each different batch has its own DNA code serving as a log of the fabrication history. The yield rate of the each batch depends on its DNA code.
  • FIG. 1 is a flow diagram of fabrication processes for one batch passing several fabrication machines.
  • each circle represents one element code of the fabrication machines.
  • a batch 10 needs to pass five fabrication processes, A, B, C, D, and E.
  • the batch 10 may be fabricated by, for example, passing a sequence of fabrication process, which is, for example, represented by a DNA code of A 1 -B 1 -C 2 -D 3 -E 2 .
  • Each of A 1 , B 1 , C 2 , D 3 , and E 2 is one DNA element code, representing a certain fabrication machine employed by a certain fabrication process.
  • FIG. 2 is an AND logic operation algorithm of an analysis method for a consequent effect, according to a preferred embodiment of the invention.
  • FIG. 2 for example, there are eight fabrication process necessarily to be passed for one complete fabrication, and there are three available machines for each fabrication process.
  • the second row shows one high-yield DNA code, which is one of fabrication machine clusters, having relatively higher yield rate.
  • the third row that is, a low-yield DNA code
  • a 1 -B 2 -C 2 -D 3 -E 2 -B 3 -D 2 -A 2 represents a DNA code with relatively lower yield rate, and is to be analyzed by the method of the invention.
  • the fabrication process may include, for example, a first stage 20 and a second stage 22 .
  • the second stage 22 may also repeatedly use a previous fabrication machine.
  • the logic operation algorithm is, for example, following.
  • An AND operation for example, is performed on each DNA element code between the high-yield DNA code and the low-yield DNA code. If the DNA element codes of the high-yield DNA code and the low-yield DNA code in the same process are the same, an AND logic operation result is indicated by “1”, and otherwise is indicated by “0”.
  • the second, sixth, and the seventh fabrication process have “0” logic results because the pairs (B 1 , B 2 ), (B 1 , B 3 ), and (D 1 , D 2 ) include different machines in their individual process. All of the fabrication machines, B 2 , B 3 , and D 2 may cause a poor yield rate.
  • B 2 in the low-yield DNA code at the second process can be replaced by either B 1 or B 3
  • B 3 in the low-yield DNA code at the sixth process can be replaced by either B 1 or B 2
  • D 2 in the low-yield DNA code at the seventh process can be replaced by either D 1 or D 3 .
  • the yield rate can be increased. Fabrication time and fabrication cost can be also reduced.
  • the analysis of the low-yield DNA codes can be more freely performed with more choices of the referencing high-yield DNA code.
  • Several low-yield DNA codes can be analyzed in one time of analysis. The analysis can also be performed beforehand if the yield rate has a tendency to be decreasing. Therefore a consequent effect of a fabrication machine cluster can be judged beforehand. Since the analysis algorithm is globally applied and automatically performed, a large amount of data can be analyzed without an ignorance of any possible reason causing the poor yield rate.
  • the DNA code can also represent any kind of a cycling series of processes or machines.
  • the invention has following characteristics.
  • the analysis method using a logic analysis algorithm to analyze the consequent effect of the fabrication machine cluster or a discrepancy between each of same-type fabrication machines.
  • the analysis can be performed automatically with a large amount of data or the DNA codes so that there is no missing factor, which may potentially cause a fabrication failure, resulting in the poor yield rate.

Abstract

An analysis method for analyzing a consequent effect of a fabrication machine cluster is provided. The method uses an AND logic algorithm to compare a fabrication machine cluster having a low yield rate with a fabrication machine cluster having a high yield rate. Each fabrication machine cluster includes a series of machines. Each of machine is used in a fabrication process. If an element machine used in a process for the fabrication machine cluster having a low yield rate is different from an element machine used in the same process for the fabrication machine cluster having a high yield rate, a logic result is indicated as “0”. Otherwise, the logic result is indicated as “1”. A modification of the fabrication cluster may be done by replacing the machine indicated by “0” with other available machine.

Description

    CROSS-REFERENCE TO RELATED APPLICATION
  • This application claims the priority benefit of Taiwan application serial no. 87115637, filed Sep. 19, 1998, the full disclosure of which is incorporated herein by reference. [0001]
    • BACKGROUND OF THE INVENTION
  • 1. Field of the Invention [0002]
  • This invention relates to semiconductor fabrication, and more particularly to a method using a logic algorithm for analyzing a consequent effect of a fabrication machine cluster. [0003]
  • 2. Description of Related Art [0004]
  • In semiconductor fabrication, one batch, which includes several wafers, needs to pass several fabrication machines in different fabrication processes, such as deposition, etching, or sputtering machines. Each fabrication machine has its own performance and its own series number. In order to have massive production, a same fabrication process may use same-type fabrication machines provided by different manufacturers. Since semiconductor fabrication needs very high precision and highly reduced dimension, discrepancy of fabrication performance between each of the same-type different machines is strictly required to be consistent in a certain small range. In addition, for one fabrication machine, its fabrication performance may also drift away after a period of operation time. All above fabrication variances may cause a fabrication failure due to a consequent effect, which is an accumulated effect of the fabrication variances in each machine used for each fabrication process. The yield rate is therefore reduced. A poor yield rate may particularly occurs at a certain fabrication machine cluster, which includes several machines sequentially used in each step of fabrication processes. Each machine has its own series number to be recorded for the fabrication machine cluster. [0005]
  • If manufacturers want to analyze the reason that causes the poor yield rate on that certain fabrication machine cluster, conventionally, it is done by manufacturers using their experiences. If the fabrication machine cluster includes a complicate situation or a large amount of data, out of handle by the manufacturers, to be analyzed, some important key reasons may often be ignored, resulting in a wrong judgement. [0006]
  • Moreover, a conventional method for analyzing discrepancies between different machines used in a same fabrication process is not suitable for an analysis of the consequent effect in a fabrication machine cluster. The actual reasons causing the poor yield rate still cannot be analyzed out. [0007]
  • In order to increase a total yield rate, an analysis method to effectively analyze the consequent effect is therefore strongly desired to more precisely judge the reasons causing the poor yield rate in certain fabrication machine clusters. [0008]
    • SUMMARY OF THE INVENTION
  • It is therefore an objective of the present invention to provide an analysis method for analyzing a consequent effect of a fabrication cluster. The analysis method can handle a large amount of data for a complicate fabrication cluster so that an actual reason causing a poor yield rate in a certain fabrication machine cluster can be precisely determined. A modification of on the fabrication machine cluster having low yield rate can be precisely done to increase its yield rate. [0009]
  • In accordance with the foregoing and other objectives of the present invention, an analysis method for analyzing a consequent effect of a fabrication cluster is provided. The analysis method includes providing a data base of high-yield DNA codes. A DNA code represents a sequential machine number of a fabrication machine cluster. The data base of the high-yield DNA codes include several DNA codes, all of which have high yield rate. A low-yield DNA code is also provided to be analyzed. An “AND” logic operation preferably is performed to compare the low-yield DNA code with one of the high-yield DNA codes from the data base. Each DNA element code in the DNA codes represents one used fabrication machine in a specified fabrication process. If a DNA element code of the low-yield DNA code is consistent with a DNA element code of the high-yield DNA code, a resulting element code is indicate by “1” for this specified fabrication machine. If they are different, this resulting element code is indicated by “0”. After the low-yield DNA code is completely compared, for example, at least one of the resulting element code is indicated by “0”. A modification process is performed to replace the resulting element having “0” with one of other same-type machines in its specified fabrication process. [0010]
  • In the foregoing, the analysis method of the invention for analyzing the consequent effect is using a logic algorithm to determine fabrication machines, which may cause the poor yield rate. The analysis method can analyze a large amount of data with a complicate fabrication machine cluster without missing any possible factor causing the poor yield rate.[0011]
    • BRIEF DESCRIPTION OF DRAWINGS
  • The invention can be more fully understood by reading the following detailed description of the preferred embodiment, with reference made to the accompanying drawings as follows: [0012]
  • FIG. 1 is a flow diagram of fabrication processes for one batch passing several fabrication machines; and [0013]
  • FIG. 2 is an AND logic operation algorithm of an analysis method for a consequent effect, according to a preferred embodiment of the invention.[0014]
    • DETAILED DESCRIPTION OF PREFERRED EMBODIMENT
  • In fabrication processes, each batch has it own log including a history of the batch performed in a certain fabrication machine at a certain step. Each machine is indicated by, for example, an element number or an element code. All the sequentially used fabrication machines form a fabrication machine cluster. The fabrication machine cluster contains those element codes, which form a DNA code like a biologic DNA structure. One DNA code represents one fabrication machine cluster or one history of the batch passing fabrication machines in the whole fabrication process. Each different batch has its own DNA code serving as a log of the fabrication history. The yield rate of the each batch depends on its DNA code. [0015]
  • For example, FIG. 1 is a flow diagram of fabrication processes for one batch passing several fabrication machines. In FIG. 1, each circle represents one element code of the fabrication machines. A [0016] batch 10 needs to pass five fabrication processes, A, B, C, D, and E. There are three same-type fabrication machines, indicated by 1, 2, and 3, available for each process. Each of these three same-type fabrication machines may also be provided by different machine manufacturers. The batch 10 may be fabricated by, for example, passing a sequence of fabrication process, which is, for example, represented by a DNA code of A1-B1-C2-D3-E2. Each of A1, B1, C2, D3, and E2 is one DNA element code, representing a certain fabrication machine employed by a certain fabrication process.
  • FIG. 2 is an AND logic operation algorithm of an analysis method for a consequent effect, according to a preferred embodiment of the invention. In FIG. 2, for example, there are eight fabrication process necessarily to be passed for one complete fabrication, and there are three available machines for each fabrication process. The second row shows one high-yield DNA code, which is one of fabrication machine clusters, having relatively higher yield rate. There are several high-yield DNA codes stored in a data base. The one, that is, A[0017] 1-B1-C2-D3-E2-B1-D1-A2 is one of the data base of the high-yield DNA codes. The third row, that is, a low-yield DNA code, A1-B2-C2-D3-E2-B3-D2-A2 represents a DNA code with relatively lower yield rate, and is to be analyzed by the method of the invention. The fabrication process may include, for example, a first stage 20 and a second stage 22. The second stage 22 may also repeatedly use a previous fabrication machine.
  • The logic operation algorithm is, for example, following. An AND operation, for example, is performed on each DNA element code between the high-yield DNA code and the low-yield DNA code. If the DNA element codes of the high-yield DNA code and the low-yield DNA code in the same process are the same, an AND logic operation result is indicated by “1”, and otherwise is indicated by “0”. In the DNA codes shown in FIG. 2, the second, sixth, and the seventh fabrication process have “0” logic results because the pairs (B[0018] 1, B2), (B1, B3), and (D1, D2) include different machines in their individual process. All of the fabrication machines, B2, B3, and D2 may cause a poor yield rate. Manufacturers can replace the fabrication machines, B2, B3, and D2 with the other two available fabrication machines. For example, B2 in the low-yield DNA code at the second process can be replaced by either B1 or B3; B3 in the low-yield DNA code at the sixth process can be replaced by either B1 or B2; and D2 in the low-yield DNA code at the seventh process can be replaced by either D1 or D3.
  • After modifications of the low-yield DNA code, the yield rate can be increased. Fabrication time and fabrication cost can be also reduced. [0019]
  • Since a sufficiently large number of high-yield DNA codes can be collected to form a data base, the analysis of the low-yield DNA codes can be more freely performed with more choices of the referencing high-yield DNA code. Several low-yield DNA codes can be analyzed in one time of analysis. The analysis can also be performed beforehand if the yield rate has a tendency to be decreasing. Therefore a consequent effect of a fabrication machine cluster can be judged beforehand. Since the analysis algorithm is globally applied and automatically performed, a large amount of data can be analyzed without an ignorance of any possible reason causing the poor yield rate. In the above descriptions, the DNA code can also represent any kind of a cycling series of processes or machines. [0020]
  • In conclusion, the invention has following characteristics. The analysis method using a logic analysis algorithm to analyze the consequent effect of the fabrication machine cluster or a discrepancy between each of same-type fabrication machines. The analysis can be performed automatically with a large amount of data or the DNA codes so that there is no missing factor, which may potentially cause a fabrication failure, resulting in the poor yield rate. [0021]
  • The invention has been described using an exemplary preferred embodiment. However, it is to be understood that the scope of the invention is not limited to the disclosed embodiment. On the contrary, it is intended to cover various modifications and similar arrangements. The scope of the claims, therefore, should be accorded the broadest interpretation so as to encompass all such modifications and similar arrangements. [0022]

Claims (10)

What is claimed is:
1. An analysis method for analyzing a consequent effect of a fabrication machine cluster, the analysis method comprising:
providing a data base comprising a plurality of high-yield DNA codes and a low-yield DNA code, wherein each DNA code comprises a plurality of DNA element codes, each of the DNA element codes represents a machine employed by a process, and each the process comprises a plurality of available machines, that are, available DNA element codes;
comparing the low-yield DNA code with one of the high-yield DNA codes from the data base, wherein if any of the DNA element codes of the low-yield DNA code is different the DNA element code of the high-yield DNA code in the same process, the any of the DNA element code of the low-yield DNA code is indicated to form at least one indicated DNA element code; and
modifying the indicated DNA element code of the low-yield DNA code with one of the other available DNA element codes used in the same process.
2. The analysis method of claim 1, wherein the machine comprises a semiconductor fabrication machine used in a semiconductor fabrication process.
3. The analysis method of claim 1, wherein the step of modifying the indicated DNA element code comprises replacing the indicated DNA element code of the low-yield DNA code with one of the DNA element code of the high-yield DNA code belonging to the same process.
4. The analysis method of claim 1, wherein the step of comparing the low-yield DNA code with one of the high-yield DNA codes comprises a logic algorithm.
5. The analysis method of claim 1, wherein the step of comparing the low-yield DNA code with one of the high-yield DNA codes comprises an AND logic algorithm for comparison, in which the indicated DNA element code is indicated by “0”, and the DNA element codes of the low-yield DNA code other than the indicated DNA element code are indicated by “1”.
6. An analysis method for analyzing a consequent effect of a fabrication machine cluster, the analysis method comprising:
providing a high-yield DNA code and a low-yield DNA code, wherein each DNA code comprises a plurality of DNA element codes in a series order, each of the DNA element codes represents a machine employed by a process, and one the process comprises a plurality of available machines, that are, available DNA element codes;
operating a logic operation between the high-yield DNA code and the low-yield DNA code to indicate the DNA element codes of the low-yield DNA code, wherein each of the DNA element codes of the low-yield DNA code is different from an operated one of the DNA element codes of the high-yield DNA, then the each of the DNA element codes of the low-yield DNA code is indicated by “0”, and otherwise is indicated by “1”; and
modifying the each of the DNA element codes indicated by “0” in the low-yield DNA code.
7. The analysis method of claim 6, wherein the machine comprises a semiconductor fabrication machine used in a semiconductor fabrication process.
8. The analysis method of claim 6, wherein in the step of modifying the each of the DNA element codes indicated by “0”, the each of the DNA element codes indicated by “0” is replaced by other one of the available DNA codes.
9. The analysis method of claim 6, wherein in the step of operating a logic operation, the logic operation comprises an AND logic operation.
10. The analysis method of claim 6, wherein the DNA element code comprises an encoded data, which carries sufficient informs to identify the machine itself and the process, which employs the machine.
US09/200,330 1998-09-19 1998-11-25 Method for analyzing a fabrication machine cluster using-genetic algorithms Abandoned US20020026298A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
TW87115637 1998-09-19
TW087115637A TW451287B (en) 1998-09-19 1998-09-19 Succession effect analysis method of production machine cluster

Publications (1)

Publication Number Publication Date
US20020026298A1 true US20020026298A1 (en) 2002-02-28

Family

ID=21631413

Family Applications (1)

Application Number Title Priority Date Filing Date
US09/200,330 Abandoned US20020026298A1 (en) 1998-09-19 1998-11-25 Method for analyzing a fabrication machine cluster using-genetic algorithms

Country Status (2)

Country Link
US (1) US20020026298A1 (en)
TW (1) TW451287B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050197872A1 (en) * 2004-03-08 2005-09-08 Sap Aktiengesellschaft System and method for performing assortment planning
WO2006016827A1 (en) * 2004-08-09 2006-02-16 Abb Sp. Z O. O. A method and a device for diagnosing technical equipment

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050197872A1 (en) * 2004-03-08 2005-09-08 Sap Aktiengesellschaft System and method for performing assortment planning
WO2006016827A1 (en) * 2004-08-09 2006-02-16 Abb Sp. Z O. O. A method and a device for diagnosing technical equipment
JP2008509420A (en) * 2004-08-09 2008-03-27 エービービー・エスピー．・ゼット．オー．オー． Method and device for diagnosing technical equipment
KR100915720B1 (en) * 2004-08-09 2009-09-04 에이비비 리써치 리미티드 A method for diagnosing technical equipment

Also Published As

Publication number Publication date
TW451287B (en) 2001-08-21

Similar Documents

Publication Publication Date Title
US6415395B1 (en) Method and system for processing repair data and fault log data to facilitate diagnostics
US6952657B2 (en) Industrial process fault detection using principal component analysis
KR100986789B1 (en) Method and apparatus for monitoring tool performance
US6625589B1 (en) Method for adaptive threshold computation for time and frequency based anomalous feature identification in fault log data
US7197414B2 (en) System and method for identifying a manufacturing tool causing a fault
KR100873114B1 (en) Method and apparatus for automatic sensor installation
Tuson et al. Adapting operator settings in genetic algorithms
Lu et al. Sub‐PCA modeling and on‐line monitoring strategy for batch processes
US6862484B2 (en) Controlling method for manufacturing process
US5923553A (en) Method for controlling a semiconductor manufacturing process by failure analysis feedback
US7974728B2 (en) System for extraction of key process parameters from fault detection classification to enable wafer prediction
Wang et al. Exact variable-length anomaly detection algorithm for univariate and multivariate time series
US6980873B2 (en) System and method for real-time fault detection, classification, and correction in a semiconductor manufacturing environment
KR100503771B1 (en) Method of and apparatus for extracting abnormal factors in a processing operation
US7583833B2 (en) Method and apparatus for manufacturing data indexing
US20020055801A1 (en) Fault detection and virtual sensor methods for tool fault monitoring
US6922600B1 (en) System and method for optimizing manufacturing processes using real time partitioned process capability analysis
Susto et al. An adaptive machine learning decision system for flexible predictive maintenance
US20050283498A1 (en) System and method to build, retrieve and track information in a knowledge database for trouble shooting purposes
Weiss et al. Improving quality control by early prediction of manufacturing outcomes
US7212950B2 (en) Methods and apparatus for equipment matching and characterization
US20020026298A1 (en) Method for analyzing a fabrication machine cluster using-genetic algorithms
US7031860B2 (en) Method and system of semiconductor fabrication fault analysis
US20040138856A1 (en) Method for analyzing final test parameters
CN114556384A (en) Machine learning variable selection and root cause discovery through cumulative prediction

Legal Events

Date Code Title Description
AS Assignment

Owner name: UNITED SEMICONDUCTOR CORP., TAIWAN

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:YEH, MENG-LIN;WU, JASMINE;REEL/FRAME:009613/0875

Effective date: 19981026

AS Assignment

Owner name: UNITED MICROELECTRONICS CORP., TAIWAN

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:UNITED SEMICONDUCTOR CORP.;REEL/FRAME:010579/0570

Effective date: 19991230

STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION